The paper presents a short summary of the state of the art with respect to fatigue strength of reinforced and prestressed concrete structures. The simplification made in ordinary design evaluations are shortly discussed. Offshore concrete structures are exposed to an environment which is different from that compared with ordinary land structures. The special feasures of this difference will be discussed. Specially, the nature of the environmental loads causes a random load application in the structure. The paper will discuss how this may be tackled in design. Different methods used in evaluation the fatigue strength of offshore concrete structures are discussed and a proposal is made, on how to formulate a design criteria for offshore concrete structures with respect to fatigue strength evaluation.

This paper presents results from an experimental investigation in which thirteen rectangular reinforced concrete beams were subjected to cyclic loading within service conditions. One beam was designed to measure the variation in steel stress and bond stress along the beam length, particularly between cracks and in the anchorage zones near the beam ends. The variables in the other twelve beams included the area of concrete concentric with each tension rebar and the ratio of compression reinforcement to tension reinforcement. Resulting bond deterioration is analyzed and its effect on bond stress distribution as well as crack development is illustrated and discussed.

As usual with other materials the constant-amplitude test (Wdhler) is widely adopted as the criterion for the fatigue sensitivity of concrete. In this test the number of load repetitions Ni the material can stand before failure occurs, is determined. In general stresses vary in a more erratical and random way. As for concrete from tests no relation is known between this kind of loading and the service life of a structure, Miner's rule is adopted for predicting this life on basis of constant-amplitude tests. According to this rule failure will occur if the following condition is satisfied: where c is the number of stress cycles during the service life. As the number of stress repetitions N. is a stochastic quantity, Miner's rule will also be stochastic. An experimental verification of the rule is therefore complicated. So a theory is giventose-parate in Miner's rule the possible uncertainty in the rule from the influence of the dispersion in N. With the aid of this theory test results of about 220 variable-amplitude tests have been verified. Miner's rule proved to be very accurate to predict the life-time of the test cylinders. At the moment the test program is continued with random loading tests. The results of that will soon be available.

Results are presented from a study carried out at the Building Research Establishment, United Kingdom, on the effect of rate of‘compressive loading on the fatigue characteristics of plain concrete. Specimens made with a typical normal weight aggregrate, gravel, and a manufactured lightweight aggregate, Lytag, were tested at one of two constant rates of stressing and unstressing, 0.5 MN/m2s and 50 MN/m2s. Using as a basis the static strength of the concrete at a standard stressing rate (0.25 MN/m2s), two distinct S-N curves were obtained for each type of.concrete. When related to the static strength at the same stressing rate as in the corresponding fatigue test, the results fell onto a single curve for each type of concrete. In studying the development of axial strains in the concrete during a fatigue test, it was found that the rate of strain increase per cycle of load was constant for most of the life of the specimen. It is shown that a strong correlation exists between this strain rate and the fatigue life of the specimen, and that this can be used to determine relationships between the strain rate and the true level of loading. Using these re1ationships'S-N curves are produced with very little scatter of results.

Reinforced concrete beams of different sizes and reinforcement arrangements were repeatedly loaded at frequencies of 0.5 and 1 Hz. Curvature and deflection are presented as ratios of the initial values and a unique relation is obtained between each of these ratios and the logarithm of the number of load repetitions. Each ratio increases gradually in a curved form to about 0.5 x l06 repetitions from where it takes a straight line form with a slope varying between 0.2 and 0.225. The long term deflections of beams tested in earlier investigations are predicted with good agreement. A similar expression is given for maximum crack width and predicted long term values are in reasonable agreement with those measured by earlier investigators.